中文摘要：

为了进一步提高差分演化算法的性能，提出一种采用保存基因的2-Opt一般反向差分演化算法，并把它应用于函数优化问题中。新算法具有以下特征：（1）采用保存被选择个体基因的方式组成参加演化的新个体。保存基因的方法可以很好的保持种群多样性；（2）采用一般反向学习（GOBL）机制进行初始化，提高了初始化效率；（3）采用2-Opt算法加速差分演化算法的收敛速度，提高搜索效率。通过测试函数的实验，并与其他差分演化算法进行比较。实验结果证实了新算法的高效性，通用性和稳健性。

英文摘要：

To improve the performance of differential evolution （ DE ）, a novel DE named generalized opposition-and-2-Opt-based dif- ferential evolution algorithm with reserved genes （2-Opt-GO-RGDE） is proposed in the paper, and then it is used to solve the func- tion optimization problems. The new algorithm has the following characteristics： （ 1 ） The new individuals can be produced by the combination of genes of the selected chromosomes. These new individuals are evolved with other individuals in the population. It can maintain the diversity of population; （2） It applies generalized opposition-based learning （ GOBL） strategy to generate initial popula- tion, and this initialization method improves the efficiency of the initialization; （ 3 ） It applies 2-Opt algorithms to accelerate DE, and this method improves the efficiency of search. Experiments are used to compare the MMT-ODE with other algorithms. The results show that 2-Opt-GO-RGDE keeps the most rapid convergence rate of all techniques and obtains the global optima for most benchmark problems.